From Google searches, to spam filters, auto text recommendations, similar image suggestions, and Amazon’s Alexa voice requests, Artificial Intelligence (AI) is becoming part of many people’s everyday lives. But how can AI contribute to creating a positive impact for those in need of essential resources and international support?

AI for agriculture

• Crop yield prediction – Using satellite or drone images AI can quickly determine which crops are being planted, yield estimations and provide location details. This can help with streamlining agricultural markets and improving farming techniques. 
• Crop disease identification – Crop image recognition can be used to find and mitigate risks more quickly such as crop diseases, insect infestations, droughts, and floods. 
• Vertical farming – In the future, “vertical farming” will become increasingly economical. There are several companies that already grow crops in huge warehouses stacking many layers of crops. AI then measures growth and yield rates, and automatically adjusts lighting, water, humidity, temperature, and other factors to optimize for yield. Some vertical farming companies claim to use 95% less water, with more than 10 times the yield of traditional farms per square foot.

AI for healthcare 

• Diagnosing illness – AI is famously being used to “read” everything from MRIs to PET scans to echocardiograms, and dozens of other scans. AI can already classify heart images that show heart disease more accurately than doctors can (92% accuracy for AI versus 79% for doctors).
• Identifying genetic anomalies – With the cost of genetic analysis dropping sharply, it will be accessible in the developing world in the coming decades. AI can be used to look for genetic anomalies, cancer, and other genetic diseases. Understanding the genetic basis can help with the prevention and mitigation of these kinds of diseases.
• Determining how diseases spread – AI can analyze vast social networks to understand how diseases have spread, and predict how they will spread. This can help prevent the spread of disease, and understand where diseases are likely to spread.
• Finding cures – Recently, a research team fed an AI program thousands of research papers. Based on that data, they then asked the program to predict what words would come next when they started sentences. The machine found relationships and new proposed solutions that had never been seen before. Similar applications will help modern medicine to find new cures.

AI for government services

• Automating repetitive tasks – Tasks such as checking documents for accuracy, automated speeding tickets, or ensuring compliance with registration and reporting requirements have the capacity to increase efficiency and quality control. 
• Fraud detection – New systems to look for anomalous data in taxes, invoices or financial records.
• Improved service delivery – Predicting who will need what services based on demographics, locations, poverty and other factors.
• Counting populations – Using AI trained in image and pattern recognition to provide more accurate estimates of the populations of slums, cities and rural areas.
• Estimating service needs – Using pattern recognition to determine whether a neighborhood is rich or poor and estimating service needs.
• Predicting disasters – Using past & current data to better predict floods, droughts, storms, and even earthquakes, and help mitigate the impacts to life and property.  
• Determining the scope of disasters – AI could scan social media images after a disaster to determine the scope, scale and impact, making it easier to deploy the correct level of response and assistance to the area.

AI for uncovering potential criminal behavior

• Finding human trafficking victims – based on image recognition 
• Predicting poaching – Using pattern detection to better understand where and when poaching occurs, and by whom. Once these details are better understood, resources can be better deployed to stop poaching.

• Uncovering tax fraud – By looking for anomalous data and incongruencies in taxes, AI can spot potential fraud much more easily than humans (who find the tasks repetitive and often boring).
• Uncovering terrorist networks – Analyzing communication patterns, looking for similarities with other networks, reviewing financial transactions or other data usage related to terrorism.

AI for improving communication accuracy

• Identify Fake News – AI is already being to determine the accuracy of some news items posted to social media.
• Identify False Actors – By looking for similarities across multiple social media accounts, looking at how they behave online, and other factors. 
• Identifying Fake Videos – by using image recognition to spot erroneous movements, slowed down or sped-up videos, or other fakes.

Keeping a level head: the downsides of AI